Dynamic Time Lag Regression: Predicting What & When
Authors: Mandar Chandorkar, Cyril Furtlehner, Bala Poduval, Enrico Camporeale, Michele Sebag
ICLR 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | A proof of concept on synthetic problems is presented. Finally, the empirical results on the solar wind modelling task improve on the state of the art in solar wind forecasting. |
| Researcher Affiliation | Academia | Mandar Chandorkar Centrum Wiskunde en Informatica Amsterdam 1098XG Cyril Furtlehner INRIA-Saclay Bala Poduval University of New Hampshire Durham, NH 03824 Enrico Camporeale CIRES, University of Colorado Boulder, CO Michèle Sebag CNRS Univ. Paris-Saclay |
| Pseudocode | Yes | Algorithm 1: DTLR algorithm |
| Open Source Code | No | The algorithm code is available in supplementary material and will be made public after the reviewing period. |
| Open Datasets | Yes | The cause series xt includes the solar magnetic ... produced by the Current Sheet Source Surface model (Zhao and Hoeksema, 1995; Poduval and Zhao, 2014; Poduval, 2016)... The effect series, the hourly solar wind data is available from the OMNI data base from the Space Physics Data Facility 5. Footnote 5 provides the URL https://omniweb.gsfc.nasa.gov. |
| Dataset Splits | Yes | DTLR is validated using a nine fold cross-validation (Table 3 in appendix), where each fold is a continuous period corresponding to a solar rotation. |
| Hardware Specification | No | The paper does not specify any hardware details such as CPU or GPU models used for experiments. |
| Software Dependencies | No | The paper mentions neural network architectures and activation functions (Re Lu, Sigmoid) but does not provide specific version numbers for any software libraries or frameworks used. |
| Experiment Setup | Yes | The initialization of hyper-parameters α and σ is settled using preliminary experiments (same setting for all considered problems: α U(0.75, 2); σ2 U(10 5, 5))." and "Table 4: Network Architecture Details ... Layer sizes Activations ... I 2 [40, 40] [Re LU, Sigmoid] |